The preliminary release of the complete true rankings are now available online. You can check out:

http://dragonboat-truerank.blogspot.com/

The ranking is preliminary as the results still need to be reviewed and I'm aware of a few small changes that need to be completed. If you have questions regarding the ranking or a teams placement in the ranking you can now contact:

truerankings@hotmail.com

The next step for the rankings will be to add some new festivals if possible along with some cleanup type details. The methodology will be posted when I get a chance. If you are interested in helping out feel free to drop me a line.

Cheers, I hope everyone enjoys. The top 250 teams to compete in either Toronto, Pickering, Sudbury, Ottawa, Lachine, or Montreal are included.

Power Demon

1) I need to include some races for the Miracle Dragons. I had a different spelling of their name but this will be fixed shortly.

2) The X-pads fall considerably because (I believe) of the results from Lachine. I need to move over a file from a laptop and then I will be able to answer festival specific questions for teams.

3) There are a lot of teams really close together!!!!!!!!!!!!!!!!!!!!

If you look strictly at the rankings and note that a team you beat in one race is like twenty places ahead of you, make sure you look at their average time. They may only have an average time of .5 seconds less than you. These rankings include all races, and though I belive they are good, are not perfect. It is quite possible that estimating error could move you up or down. I would guess that anyone within a second of you would be fair game to take down if you have a good race.

Power Demon, I want to marry you!

Thanks! Great job.

Doesn't look right in Firefox though, Blogger's about box is overlayed over times column. In IE, the same box is at the bottom, after the table.
Not a biggie, just an FYI.
Again, thanks for all the hard work figuring this out.

PD, this really looks good! I was a big skeptic to begin with, but with every new iteration this is looking better and better.

One problem is evident though. One hit wonder teams come out looking really good, but they only did one regatta. There is no demonstration of consistency. Could they muster that roster twice? Can they go that fast again?

Some are no-brainers. For instance, I'm sure MCC/SRS could do it. But Events Alive is AWOL after Ottawa, their own backyard. Maybe if they went out of town, they lose half their A-list paddlers and drop 3 seconds off the pace? Not saying that IS the case, just that it should be considered.

Maybe there should be a minimum race/regatta limit?

to the last poster ... its just an average to get an idea ... i'm sure every team on the list hasn't paddled with the SAME roster every race ... most teams have subs/ringers once in a while ...

as for 1 hit wonders, there's a reason why they're called that :P once more data comes out, the ranking will look even better

Regards to firefox, yeah, the side menu is ontop of the list ... kinda annoying but still good!

Very nice work indeed!

Thanks for the effort you put into this.

2 questions:

1- Are there available rankings for all teams or just top 250?

2- How many teams total qualified to be ranked in your system? Would be nice to see how many teams total there are to give a sence of perspective to the top 25 list.

Great work! The moderators of this forum should link http://dragonboat-truerank.blogspot.com/ instead of the rankings site they have now which is never updated

Did you filter out the NY teams? I can't find MAD but I see DCH

Does this include the Hamilton and London races? Where does Swordfish rank?

Are there available rankings for all teams or just top 250?

just 250?

Where's NY Wall Street?

I'm really glad people appreciate my work and I'm really happy that people are finding this a useful ranking system. In answer to questions:

1) Hopefully the formatting problems were fixed this morning. Please let me know if any changes still need to be made.

2) One Hit Wonders: This ranking tries to get at "if a team showed up with their average crew, how fast would they likely go for an average race". In my mind, if a team like Events Alive had the same roster, the fact that they went as fast as they did for all 4 races in Ottawa is enough to signal that it wasn't a fluky race, fast lane, or improperly timed showing. In essence, the ranking boils down to "who was fastest in 2005 if we could hypothetically line up all the teams with their average roster for one race". I honestly like the fact that this ranking doesn't reward teams for the number of races they enter like the GP ranking does. I included the number of races column to give an indication of who realistically shows up to festivals and who to look out for. Everyone knows that Events Alive doesn't show up outside of Ottawa but that doesn't make them any slower.

3)I originally had intended to keep the rankings fully Canadian and as a result filtered out the American teams (obviously missing some like DCH). The reasoning behind this was to keep teams from the U.S. that show up once every 10 years from bumping down boats one year and then disappearing from the ranking in the next. On the other hand, I think teams like DCH or MAD that do show up regularly should be included. Any thoughts on this? It will not be possible to included American teams in 2005 but maybe we can make a returning rule. If they paddled in Canada last year then they can be ranked.

4) The ranking currently holds data on roughly 500-600 teams (i.e. anyone who competed in the 6 listed festivals). As more festivals are included this number will grow. As mentioned in earlier posts, I feel there may still be a weakness in the adjustment for the slower teams in Sudbury because of low sample size. This has not yet been looked at.

I did not post the results for all teams because it would not be good sportsmanship to list the slowest mixed boats. I will however take team specific requests if you email truerankings@hotmail.com. Just email and ask for the results for Swordfish for example.

5) I may also release the top 10 teams for men’s and women’s when I have a chance. Just be wary that it is likely that men’s and women’s teams are often thrown together with good paddlers from other boats and really hard to keep track of.

6) As for any marriage proposals, I'm afraid after this weekend Power Demon is no longer on the market.

Cheers,

Power D

Yup, formatting's fixed.

Great work! best ranking yet

Excellent work PD

"Does this include the Hamilton and London races? Where does Swordfish rank?"

I would guess 275-300ish

"3)I originally had intended to keep the rankings fully Canadian and as a result filtered out the American teams (obviously missing some like DCH). The reasoning behind this was to keep teams from the U.S. that show up once every 10 years from bumping down boats one year and then disappearing from the ranking in the next. On the other hand, I think teams like DCH or MAD that do show up regularly should be included. Any thoughts on this? It will not be possible to included American teams in 2005 but maybe we can make a returning rule. If they paddled in Canada last year then they can be ranked."

I think any USA crew who attends a Canadian festival should be ranked it gives the rankings an added dimension. I can also see why they might not be included since some of the NYC crews would pull Ontario paddlers to fill out a crew. Crews such as those will be difficult to rank since race times may fluctuate greatly. One way maybe to overcome it is to collect data from the USA regattas.

Awesome! I hope you didn't obsess over this during your whole vacation!

(one tiny thing I noticed: ops pdb.collective and ops.pdb.collective are listed as 76 and 101 respectively but they're probably the same guys.)

Dare you work in the West for a full-blown Canadian ranking?

So, in true nerdly fashion, I looked at how many teams were in each 2-second time slot. So, when it says "01 1", it means sub 2:01 time, 1 team. Or "03 3" means between 2:01 and 2:03 there were 3 teams.

01 1
03 3
05 6
07 6
09 6
11 10
13 14
15 15
17 30
19 26
21 33
23 56
25 46+

It looks like it's really hard to break 2:09 - this is a truly elite group (A?). From 2:09 to 2:15 there are a lot more good teams (B?), but it's still not huge. But, my god, the group below 2:15 is incredibly competitive! Look how many of them there are! Of course, a statistician would just tell me that's a normal curve, so what's the big deal, but I think it's pretty cool. I'm not so sure. It's almost stepwise. It's like it takes a leap to progress from one "level" to the next.

I love it, first a detailed set of rankings, now a breakdown of groups. And Nerdgirl, you're right about that stepwise thing. As a former coach of mine is fond of saying, you have to "find the next gear". Looks like it takes finding that gear to move up a step in the rankings.

Looks like Sub 2:15 is one step, and sub 2:10 is another

"Does this include the Hamilton and London races?


London, or Stoufville can't be included. I would also guess Waterloo due to the weather.

Your right, when you look at what a team has to do to make it to the next level and much faster these "A" level teams are, you realize how hard they also work. Let's not forget that a number of teams have some over 40+ in age paddlers that still kick our royal butts!! Alot of them don't come from flatwater backgrounds either. It shows you how much these people invest in this sport.

Excellent work Nerdgirl!

Check this out, by 5 second intervals:

2:00-2:05 Events Alive (1)-Scotia Rouge (10): 10 teams, top 4 percentile
2:05-2:10 Hammerheads (11)-Isshin (24): 14 teams, 5.2% of field, top 9.2 percentile
2:10-2:15 Verdun Express (25)-Chaos (59): 35 teams, 14% of field, top 23.2 percentile
2:15-2:20 PanaCEAA (60)-D'Youville (134): 75 teams, 29.6% of field, top 52.8 percentile
2:20-2:25 Drako Community (135)-Kindred Spririts (250): 116+ teams, 46% of field

I think for most large festivals, this gives you, respectively:

A
A-consolation
B-C
D-E
F-?

This is fascinating stuff! Any way you slice it, sub-2:10 is pretty rare air!

Agree with Seat#5 - can you work in the West for the next iteration...

Nice work, and I like how people broke down the groups. It's alot easier to go from 2:25 to 2:15 (10 seconds)then it is to go from 2:15 to 2:10 (only 5 seconds.
Congrats to those who have made the jump.

Power Demon getting hitched this weekend? Where can i send my wedding gift! ... Congrats!

Power Demon getting hitched this weekend? At the finish line in Welland? (hey, marathoners do it, why not a dber?)

No wonder he wanted to get this info out. It may be the last time we hear from him.

Nice Job, PD (on the rankings) and Congratulations & Best Wishes (on the weekend ... and I so do NOT mean Welland)!

Is the ranking reseted each year? If not, is it an average of all races spanning many years or weighted, givvign more weight to recent races?

I would go either new ranking (reset) each year or weighted something like this: 55-30-15 and ditch everything bellow 3 years. Actually, a new one each year is probably much better to think of it..

I've notice a bunch of duplicates that are still hanging around on the list. I should have it cleaned up by the weekend.

The ranking will be based on a yearly rotation. I'm trying to think up a good way of updating it on the fly (for example the Lachine results from 2006 replace the Lachine results from 2005 as the season progresses). I really don't know if this will be possible, but if not we'll keep it the same as this year but try to produce it in a more timely manor.

Another idea a teamate of mine had was to also publish an active ranking that only includes the average time for a team from the last festival they attended. In order to do this well though I will have to get a few more festivals into the mix. At the very least I can likely do an arrow (like the power rankings) to indicate which direction the team moved on their last festival.

All in good time....

Power D

Great stuff indeed. Maybe you can toss in Welland sport (8 teams?) as a quickie to give a quick heads up on where Piranhas rank before NACCC? It's the only important team missing.

X-Pads don't look like they'll ever fit in without some serious filtering. Their times have been all over the place all year. Oh well, the system seems to work pretty well for everyone else.

Noticed a curious clumping of the eastern teams in the 20s (2:09-2:10) before Ontario crews show up again. I'm wondering if Montreal got adjusted a little faster than necessary. Maybe it will blend a little more after GWN?

"Noticed a curious clumping of the eastern teams in the 20s (2:09-2:10) before Ontario crews show up again. I'm wondering if Montreal got adjusted a little faster than necessary. Maybe it will blend a little more after GWN?"

I also noticed that. Some of the Montreal teams were pretty high based on MIDBRF results. My guess is some T.O. teams (perhaps X-pads?) had a relatively low performance in either Lachine or Ottawa which gave a strong adjustment to the East.

Anyway PD, you know you've done something right when even people on this forum like it!

Hey PD, Really liked your rankings. Alot of work, no? Alot of time on your hands between coaching, working etc... (just kidding)

KUDOS!

U-DA-MAN

These are great and a huge data resource.........

Please keep it current if you can, it's great and no one can run and hide or bitch about stuff.... just the facts baby!

I will look more closely at the adjustments for festivals when I do up the final rankings. However, I believe that the clumpings of Toronto crews high and Montreal crews following up is correct.

The only really top top teams coming from the East are Events, Slayers, Verdun moving up and Bytown now moved down. I don't find it surprising that all the other remaining top crews are GTA. I don't belive these teams are adjusted too high. Then we have a bunch of competative but not elite crews remaining in the Montreal area. It really doesn't look funny to me. I really don't think these teams are adjusted too low. Looking at the list all the Montreal crews are very good but not yet at the elite level. A good check on this would be observing how many Toronto crews average times within a second of the MOFOS in the GTA. If a bunch of teams are generally tying with the MOFOS then it may suggest Montreal is too fast or slow.

The regressions behind the adjustment are very strong for Montreal, Lachine, Pickering and Toronto. All of these festivals have roughly 30 observations. The same holds true for the adjustment across divisions. Even if the X-pads were included in the same they would adjust the regressions very little (of note they were not used as I originally did not pick up that they were in Lachine). If any festivals are off, they are likely to be Sudbury or Ottawa, though by judgement they look fairly good.

One thing to note:

IF YOU ARE FROM GTA, you will likely take a look at the East boats and say who are these guys they are ranked too high. This is simply a function of not racing them.

IF YOU ARE FROM MONTREAL OR OTTAWA, you will likely take a look at the West boats and say who are these guys that are ranked too high. This is simply a function of not racing them.

This is just the nature of the beast. Hopefully more cross divisional racing will change this. I really don't think there is a strong bias, just there seems to be more drive for elite crews coming out of Toronto.

Power D

I must admit that your ranking is growing on me. I initially didn't really think you could get a very good gauge by time adjustments across festivals due to varying conditions from one heat to the next and lane variations, but with such a large sampling it sure seems like your rankings are pretty much where you would expect teams to end up.

Sure you can look and see that some teams haven't beaten other teams that they are ranked ahead of, but as you said, their average times are just better and that's what this is all about. Maybe the other team always has their fastest race when it counts. All good beer and peanut discussions.

What I was wondering. What I wanted to throw out, is perhaps some form of weighting to allow for more recent result to carry more weight. While this isn't such a big deal now, you will have to figure out how to carry over this years results as a basis/starting point for next years.

And even for this year, I think we would all agree that what a team did in Pickering doesn't reflect their current strength nearly as well as the upcoming NACCC and GWN races would.

Since we the season is basically 5 months (May, June, July, Aug, Sept and later) you could give the current month a 1.0 and then subtract a 0.1 weighting for every month/period in the past the results came from to a minimum 0.6. As Months/periods reappear the next year the previous periods drop off.

The weighting effects the NUMBER of races at the months average time, not the average times themselves. This is key.

So lets say a team had an average time of 3:00 in Pickering and then 2:00 at GWN (it’s an example and for arguments sake lets assume they only raced twice in each period/ month time and that the weighted value of Pickering at the end of September was 0.05 instead of 0.06).

Current calculation:
(3:00 + 3:00 + 2:00 + 2:00) / 4 = 2:30 avg time
Pickering Avg 3:00 for 2 races and GWN avg = 2:00 for 2 races

Proposed Calculation:
(((3:00 + 3:00)  * .05 weighting) + ((2:00 + 2:00) * 1.0 weighting) / resulting number of races (3 in this case)  = 2:20 avg time
(3:00 + 2:00 + 2:00) /  3 = 2:20 avg time
Pickering Avg 3:00 for 1 race (2 races * 0.5 weighting = 1) and GWN avg = 2:00 for 2 races

I think this would provide a more accurate reflection of Current strength and trends.

Anyway, great work, I am a convert and fully support your whacky obsession.

I won't lie, it's a good idea. I've been thinking myself about how to reflect a more recent rankings. It's clear that at least a handful of teams have stepped up their game and are getting weighted down by festivals earlier in the year. Another similar issue is how to update to next year. As it stands right now the ranking is an annual one for performance across 2005.

My main problem is mechanics. Right now I need to make sure that I get a sound system in place that can be updated much faster and with more festivals. I like the idea, or something similar to it and will try to get it in place as soon as possible. I might also release average times by festival if it is easy to do so such that you can see how teams performed in more detail.

To clarify:

Firstly, X-Paddling... not X-Paddlers...
And secondly, the reason that X-Paddling has times all over the map is because the pdbc mixed b crew was confused with X-Paddling (pdbc mixed a crew) in Sudbury and in Lachine. The event organizers posted the results with the wrong team names... so that's that.

Thanks for the post above. The Truerankings have now been updated to reflect the following:

1) Results from Welland
2) Correct identification of the X-Paddling Collective (though I still have the name wrong on in the rankings,..sorry).
3) Removal of 9 duplicate entries (curse access and it's case sensative ways).

The welland info may not shed a lot of light on current team performance, but it does add a few new teams to the rankings which is interesting.

A slopily written methodology is up on the true rankings site for those interested. Let me know if I can improve any areas by clarifying.

Cheers,

Power D

Welland was a great beta for the new "complete" rankings. It was the first significant chance for top East teams to mix it up with top Central teams. How did we fare (omitted are Philly and Piranhas)?

True Rankings

Mayfair
Imps
Slayers
Chiros
Scotia
Verdun
Hammerheads
Hanalei
Big Fish
Tempest
Telus

The NACCC 500m

Scotia
Mayfair
Slayers
Chiros
Hammerheads
Imps
Hanalei
Big Fish
Verdun
Telus
Tempest

Now, how do you interpret the results? Did the wheels fall off the Imps and Verdun? Did Scotia pull a huge upset?

Everyone else seems to have slotted in just about right (with the notable exception of Tempest). So far, the Rankings are holding up quite nicely. Behold the power of PD's methodology!

in reards with the last posting.... can'T say about the IMPS (I'm from the east) but Verdun did step on a wet stone on Saturday. They came back strong on the 200m where they beat the Slayers and finished 4th. Muuuuuuch better than "a" consolation on the Saturday. Telus (faster than last year) and Hyperactifs (slower than last year) were at each other's throat all week-end. Telus did better in the overall ranking (one position) and scored two victories in head to head racing on the 200m. Hyperactifs had two bad results starting on the saturday races but were strong after their last 500m on the saturday, with a chrono that would have put them in the "a" consolation.

Those two teams are not of the same caliber as Dragon Slayers, Verdun or top elite Toronto team, however, with The BMO team and Mtl Thunder, they'll give serious headaches to "b" teams of all level.


BTW : does anyone know on which website we can get the FINAL and OFFICIAL results of this week-end's racing ? Thanks a bunch !

DBC or GWN?Aren't they one and the same?

To the poster above, can you clarify who you think is a duplicate.

DBC- Means Dragon Boat Club
GWN- Means Great White North

After not making the A final, verdun put all their spares back in the boat and just made a race of it.

Who was in GWN?Ppl from Berlin?

GWN is in a week and a half, bud.

Love the rankings dude. Keep up the good work. You get my vote.

Does anyone know when the GWN teams list will be posted? I want to see who's going and test out PD's rankings

aw shucks all the USA crews are out of the rankings!

Just a little heads up. I started splitting out average times for crews by festival just to get a better look at the ranking. I'm hoping to post this up shortly.

However, by digging deaper in the ranking I uncovered a problem with Lachine results that seem to be too fast and do make sense in the context of the given methodology. I believe that I may not have recalculated the Lachine adjustment after I recalculated Montreal results using the slower boats. Once the problem is identified and corrected, boats racing in Lachine will likely rank a little lower. This is probably a mistake on my part, sorry. With any new ranking system I guess their are growing pains.

I will try to fix this and add the NACCC results before the weekend.

Cheers, Power D

Humility and Accuracy

By the way, can anyone verify that in the NACCC results:

open = mens
grand masters = mens

?

Open is generally men (could include women in theory).

Didn't notice the actual teams, but according to the bulletins they were only racing mixed category for grand masters.

You have mail PD.

Props PD.. good work

Alright Magic 8-Ball, do yer thing:

A Final
3 Predators
10 Chiros
14 Hammerheads
15 Hanalei
16 Big Fish

A Consolation
18 SCC
20 Telus
22 Immortals
26 Verdun Express
28 Mofos

B Final
29 WRCC A*
33 Driving Force
34 Wave Cutters
35 Shogun
39 Saints

B Consolation
43 Blades
44 Jetstart
45 Mayfair Warriors
46 Mojos
48 Hydrophobic

C Final
52 Alliance Evolution
57 Komodo
59 Chaos
69 Fire Dragons
71 Bod Squad

C Consolation
73 First Dragons
75 Shockwave
83 5.10
92 Aquatrain
93 Miracle Dragons

D Final
95 Torpedoes
98 Smoke on the Water
102 Banana Boat
107 Cyclones
113 Catch 22

D Consolation
137 Unisen
155 Iron Dragons
171 Fearless Feds
176 Battledragons
191 Frutz on Board

E Final
194 River Raiders
199 Stokin' Sea Monkeys
200 Sinai
206 Crime Waves
215 C Monsters

E Consolation
221 Ragin' Barracudas
222 Tea-Chin
243 Civil Dragons
250 Leaping Lizards

There are a couple of wildcards: Para Huti has to go somewhere, TAD is unranked (?), there's a team listed as "Adrenaline" but I don't know if it's Verdun, WRCC is not going full strength (mainly as juniors), the MCC team is anyone's guess.

You now know where you are fated to end up! With 50 of the top 250 teams showing up, good luck to everyone on placing higher than your ranking would put you!

Oops, gotta lookout for the GWN Women and Senior Women as well as I'm sure they'll inflict their usual damage among the mixed teams.

Don't know about that. A number of these ladies are on mixed crews that will be competing. They may not race to their full potential.

The only guess with MCC is " Open water or not ? "

Just noticed that teams that did Welland droped in average time and in the rankings. Teams that did not compete moved up by defalt. Understand slow times were due to stong head winds.
The top teams should have been under 2 minutes ,in the 1:58 range on the fast course.Instead they were 2:03+. Scotia won the finals in 2:05+. The effect of the wind is shown by the senior winners , Trout , winning their finals 2 or 3 seconds faster than Scotia.
The rankings are great but is there a way to adjust for this sort of variation in course and conditions ?

That's odd, because the whole point of these rankings is that each race IS adjusted. So unless something went flooey in PD's program, it's unlikely that all Welland teams dropped. It's possible something went flooey?

I knew that the Welland results would throw this for a loop.

Headwind was huge this year. In 2003 there was a tailwind and teams were cracking 1:58. this year the best time was 2:03.

So, using the same time-brackets as before, GWN entrants breakdown thusly (treating the event as if it included only the 49 ranked teams):

2:00-2:05 2 teams, 4% of field vs. 4% of ranking
2:05-2:10 6 teams, 12% of field vs. 5% of ranking
2:10-2:15 14 teams, 28% of field vs. 14% of ranking
2:15-2:20 13 teams, 27% of field vs. 26% of ranking
2:20-2:25 14 teams, 28% of field vs. 46% of ranking

So, what has been described as a weak field, is actually relatively strong especially among the 2nd and 3rd tier teams. As a value judgement, it looks like a very competitive event for teams in these 2 tiers to attend.

As usual, there should be a lot of great racing through each division.

My appologies for the slow update of festivals. I've been busy and closer inspection of the rankings has revealed some problems.

As pointed out all Welland teams slowed down. Which given my methodology should be impossible. Somewhere I have made a mathematical mistake and need to track it down. Also, I am having problems with the adjusted 200m races being heavily favoured for fast teams and not for slow teams. This may be because the relationship for the adjustment is no longer linear. I may try a log linear regression to see if this improves results.

My appologies. Be patient. I'd rather fully understand the data then produce something that is flawed.

MCC by open water.Funny that they have not been mentioned in A Final.

How about including the London and Woodstock races?

Maybe I missed something. I thought the regression would take care of the slower times? If everyone was slower across the board then wouldn't the adjustment bring everyone back into line with othe original results.

You cannot take times, and then simply average them to suggest rankings.
If one regatta has a considerable headwind, and you would assume that the "norm" for that event would be 2-5 seconds faster, then you have to weight the result with this in mind.
Some courses are faster, other times you have headwind/tailwind variations.

So for one race, you might suggest a factor of 0.95.
This would mean a race that finished with a time of 2:05 would result in
1:56 +/-
Do the reverse for tailwinds.

He isn't simply averaging the times. He is adjusting the times based on how much faster or slower the times are compared to other races that had the same teams.

For example if 50 teams did Pickering and Center Island he can look at those 50 teams to find out how much slower or faster Center Island was than Pickering and then use that to adjust the times.

The problem with the True Rankings is that it used Pickering as it's baseline. In fact, it seems that any results that aren't in line with Pickering are thrown out or blow the curve.

The flaw now that I see... is that the system does not adjust for possible changes/improvements in teams over the season. A team that may have started training late this year, but practiced vigourously throughout the season would probably change their positioning by the end of the year.

The True Rankings is not meant to be a predictive measure, it simply ranks teams based on past performances.

The previous poster is correct, it does mean that it doesn't take into improvements or adjustments to the teams but that doesn't necessarily mean that it's a flaw. Teams are dynamic, there will most likely be improvements in conditioning, technique, etc over the season and also changes to the roster. If there are enough changes through the season or roster, then inevitably there will be some surprises and teams that buck their place in the True Rankings for something else. That's what makes the competitions fun.

Indeed. My apologies. In my haste and exhaustion from racing (afterparty), I forgot the fundamental flaw in predictions. They predict the past, and gives an indication towards the future. Thank you for correcting me.

Man, you turn your back on the forum for like one second and suddenly misconceptions are flying left right and center!

I have determined a mathematical flaw in the rankings which is currently being fixed and will take me some time. I realised that when running my regressions I was using the "adjust from" times on the x-axis when I should have been using the "adjust to" time on the x-axis. Because the slope of the adjustment time is typically small and races close in time, this let to minor errors in most festivals. However, in festivals with a large time difference (adjusting 200m times) and for festivals with a steep slope (Ottawa) it did matter. I now have to take the time to recalculate all adjustments which will put the rankings out of commission for a bit (though I almost have NACCC and thanks to Q the woodstock festival added). If you don't have a clue what I'm talking about, well,... that's ok. The methodology on the site might help clarify. Just know the rankings will change slightly.

In responce to the above, a headwind will be caught in the regression and should not matter at all. A headwind in only one lane will matter to the extent that teams consistently race in that lane. The more races teams log the less this is likely to be relevant.

All festivals are adjusted to Toronto Island not Pickering.

Team improvement over the season is netted out of the true rankings. The regressions assume that there is an average improvement amongst teams and factors this out. As a result, improvement will only change your true ranking time if you improve more so than the boats you compete with. Think of the rankings as always mapping back to the Toronto Island festival. If teams are expected to improve as the season goes on and you don't, it suggests that if you went back and raced the Toronto festival earlier in the year you would be weaker. Times will not reflect what teams do now, but what they would be expected to do back then given their performance today.

Team improvement is only a problem for the rankings if teams in the West (Toronto) improve faster than teams in the East (Montreal) or vice-versa. Even some of this difference will be captured by the regressions if teams competate across division (as in the NACCC Welland and Montreal). This is not likely to be a large problem, but may matter more toward the end of the year.

Power D

P.S. I'll be working really hard to update so bare with me.

There were a lot of surprise results this year at GWN and it should be an interesting test of the 3 standard deviation race exclusion policy.

Off the top of my head, teams I suspect are flirting with 3 SD are Shogun, Chaos and Seamonkeys on the upside and GM, Blades and Mayfair Warriors on the downside. This may become especially obvious after times are adjusted for the headwind on Saturday.

As I suggested in my Stats Primer post, you guys might want to send PD a quick e-mail if there's something relevant you like him to take into consideration.

[BTW, PD I have total faith in you that you can tidy up any nagging technical difficulties. I <3 True Rankings. Thanks again for all your hard work.]

Was GWN 2005 as competitive as previous years?As others have stated,many top flight crews were absent due to WCCC trials in Montreal etc.Is it still going to be ranked as high as Pickering or Welland?